No problem - this: oXML.selectSingleNode("*//*[local-name()='rows']")
was an accident, or the result of a trial and error approach (whats the diff anyway - huh?:-)

There is a guy, getElementsByTagName - but I just luck into results when I use him and forget when I finish. It seems a multithreaded server side XML DOM that allows concurrent user access might someday be the data base we work with. Learning XPath seems to be an understated priority - if you hear of any good VFP books on Xpath - let me know - I prototype a lot of my queries in VFP before I commit them to a compile time engine.

There is a book - (XPath and Xpointer) - an ORiely book (me thinks) - but the notation, I remember, is not VFP notation!

We have a UT alumnus - Greg Riechert - that has ventured into the heart of XPath - but he makes us beg for mercy before he waves the wand!:-)

Is your XML file a "manifest" XML or a "tablular" xml?


>Terry,
>
>Thank you much for the additional input. I think I have a fairly complex XML file compared to the example outline you provided.
>
>The additional coding you provided does add to my arsenal for getting the data out from it.
>
>Thanks!
>John G.
>
>>In addition to Pinto's reply, yiou can address the XML in other ways that might be helpful - or might be more confusing.
>>
>>With VFP we can drive, plug AND play with an XML object until our fingers get
>>blisters on them:)
>>

*One approach might be (assuming you have an XML organized [something]
>>*like this:
>>(xml)
>>  (table}
>>    (rows)
>>      (row) . some data as additional nodes or attributes . (/row)
>>      (row) . some data as additional nodes or attributes . (/row)
>>      (row) . some data as additional nodes or attributes . (/row)
>>    (/rows)
>>  (/table)
>>
>>*Okay - let's say you're only interested in the "row" elements that are *children the (rows) element.
>>*Start your XML
>>oXML=CREATEOBJECT("MSXML2.DOMDocument.4.0") &&Try an implicit connect to 4.0
>>oXML.async=.f.
>>oXML.setProperty("SelectionLanguage","XPath")
>>oXML.load(GETFILE()) && You may need to be local to the folder
>>                     && to assure the XSD's merge
>>                     && ALso note that 4.0 will merge XSD's that
>>                     && Earlier XML objects "will not" merge.
>>* Now create a reference to the (rows) element:
>>oRows=oXML.selectSingleNode("*//*[local-name()='rows']")
>>* Ho many (row) are in (rows)
>>?oRows.length
>>* To point to a "row"
>>oRow=oRows.item(n)
>>* If the fields are stored as tags
>>oFields=oRow.childNodes
>>* If the fields are stored as attributes
>>oFields=oRow.attributes
>>* You could then address the nodenames and node texts or values using
>>* a loop, or "literally" request the node names (in either case).
>>* With attributes, some may be missing, you might need to verify they are
>>* of the parent element:
>>IF TYPE('oFields.getAttribute("txtSomeAttributeName")')==[C]
>>* returns a "True", the attribute is there.
>>* I am not sure about elements - I think I have written XMLs with an empty
>>* element "(row)(/row).
>>* You can also get a different result (a more exclusive set - fewer items)
>>* depending on your XML heirarchy (i dont know why):
>>oRow=oRows.selectNodes("*[local-name()='row']")
>>

>>There are features that associated with the tag when addressed as a "childNodes" and features that are associated when item(s) are addressed.
>>I would like the big picture - but have not fount it.
>>You may want to take note as to how the first XPath (local-name()) address the manifest as opposed to th e way local-name() is address at the bottom of the sample.
>>
>>HTH
>>
>>
>>>Thats working. I just have alot of coding ahead of me to pull the info into variables then put that into tables.
>>>
>>>Thanks,
>>>John G.
>>>
>>>>If you're using the DOM, then Intellisense is your friend. I had to process an XML file this way, and used mainly two functions: SelectSingleNode and SelectNodes.
>>>>Here are some lines that will give you an idea...

>>>>loXml = CreateObject("Microsoft.XMLDOM")
>>>>loXml.Load ( YourXMLString )
>>>>loRoot = loXML.documentElement
>>>>IF ISNULL( loRoot )
>>>>   RETURN 'Invalid XML'
>>>>ENDIF
>>>>loElement = loRoot.selectSingleNode( "CustomerName" )
>>>>lCustomerName = iif(!ISNULL(loElement),allt(loElement.Text),"" )
>>>>
>>>>lOrderNumbers = loRoot.selectNodes ("Orders/OrderNumber")
>>>>for each loElement in lOrderNumbers
>>>>   ? loElement.Text
>>>>endfor
>>>>
>>>>** you can also try loRoot.selectNodes ( "Orders/Blabla/Blablabla" )
>>>>

>>>>
>>>>I hope you get the drill :-)
>>>>Jaime
>>>>
>>>>>I've been beating my head looking for the best way to process XML files in VFP 9.0.
>>>>>
>>>>>XMLADAPTER does not work because it uses MSXML 4.0. The XML files I am getting only work with MSXML 6.0.
>>>>>
>>>>>So I was looking at using XPath to navigate through the XML. But now I'm looking at just using the DOM to do so.
>>>>>
>>>>>I am new to doing this so any help on the best direction to take and learning resources to use would be well appreciated.
>>>>>
>>>>>The XML files that will be processed are fairly complex and have many nodes. They represent purchase orders and sales orders.
>>>>>
>>>>>Thanks!